Full precision abs(src0)src1.
pow dst, src0, src1 |
---|
where
Vertex shader versions | 1_1 | 2_0 | 2_x | 2_sw | 3_0 | 3_sw |
---|---|---|---|---|---|---|
pow | x | x | x | x | x |
This instruction works as shown here.
dest = pow(abs(src0), src1);
This is a scalar instruction, therefore the source registers should have replicate swizzles to indicate which channels are used.
The scalar result is replicated to all four output channels.
This instruction could be expanded as exp(src1 * log(src0)).
Precision is not lower than 15 bits.
The dest register should be a temporary register, and should not be the same register as src1.